AI Detects Hidden Stress Markers in the Body

A groundbreaking AI tool has been developed that can detect chronic stress by analyzing adrenal gland volume from routine chest CT scans. This novel approach provides an imaging-based biomarker that correlates strongly with physiological indicators of stress, such as cortisol levels and allostatic load, and is predictive of future cardiovascular events. The implications are significant for early detection and prevention of stress-related diseases, as this method utilizes existing medical imaging data without requiring additional patient procedures.

Chronic stress is a pervasive issue in modern society, contributing to a host of physical and psychological ailments including anxiety, sleep disturbances, hypertension, and a compromised immune system. Its long-term effects are known to increase the risk of serious conditions like heart disease, depression, and obesity. Despite its widespread impact, objectively measuring chronic stress has remained a challenge, often relying on subjective questionnaires or fluctuating biochemical markers.

The research, presented at the annual meeting of the Radiological Society of North America (RSNA), introduces a deep learning AI model trained to precisely measure adrenal gland volume. Dr. Elena Ghotbi, a postdoctoral research fellow at Johns Hopkins University School of Medicine, led the development of this model. Given the millions of chest CT scans performed annually, this technology can be widely applied to evaluate the biological effects of chronic stress across diverse patient populations.

Unlike transient cortisol measurements, which offer only a snapshot of immediate stress, adrenal gland volume serves as a more stable and cumulative indicator of sustained stress exposure. This biomarker is derived from data obtained from the Multi-Ethnic Study of Atherosclerosis, which encompassed 2,842 participants. The study integrated CT scans, validated stress questionnaires, cortisol measurements, and allostatic load markers—a comprehensive dataset ideal for developing and validating an imaging-based stress biomarker.

The deep learning model retrospectively processed CT scans to segment and calculate adrenal gland volume, leading to the establishment of the Adrenal Volume Index (AVI). Researchers found a significant association between higher AVI and elevated cortisol levels, increased peak cortisol, and greater allostatic load. Individuals reporting high perceived stress also exhibited higher AVI compared to those with lower stress levels. Importantly, a direct link was observed between increased AVI and a higher left ventricular mass index, as well as an elevated risk of heart failure and mortality. Specifically, every 1 cm³/m² increase in AVI corresponded to a greater risk of adverse cardiovascular outcomes.

This innovative imaging biomarker represents a significant leap forward in understanding and quantifying the physiological burden of chronic stress. It offers a practical and scalable method to identify individuals at risk for stress-related health complications, potentially revolutionizing preventive care strategies. By leveraging routinely collected imaging data, this approach avoids the need for new tests or additional radiation exposure, making it a highly accessible and efficient diagnostic tool.